Sensitivity study of the microdosimetric kinetic model parameters for carbon ion radiotherapy

Abstract In carbon ion therapy treatment planning, the relative biological effectiveness (RBE) is accounted for by optimization of the RBE-weighted dose (biological dose). The RBE calculation methods currently applied clinically in carbon ion therapy are derived from the microdosimetric kinetic model (MKM) in Japan and the local effect model (LEM) in Europe. The input parameters of these models are based on fit to experimental data subjected to uncertainties. We therefore performed a sensitivity study of the MKM input parameters, i.e. the domain radius ( r d ), the nucleus radius ( R n ) and the parameters of the linear quadratic (LQ) model ( α x and β ). The study was performed with the FLUKA Monte Carlo code, using spread out Bragg peak (SOBP) scenarios in water and a biological dose distribution in a clinical patient case. Comparisons were done between biological doses estimated applying the MKM with parameters based on HSG cells, and with HSG parameters varied separately by ±{5, 25, 50}%. Comparisons were also done between parameter sets from different cell lines (HSG, V79, CHO and T1), as well as versions of the LEM. Of the parameters, r d had the largest impact on the biological dose distribution, especially on the absolute dose values. Increasing this parameter by 25% decreased the biological dose level at the center of a 3 Gy(RBE) SOBP by 14%. Variations in R n only influenced the biological dose distribution towards the particle range, and variations in α x resulted in minor changes in the biological dose, with an increasing impact towards the particle range. β had the overall smallest influence on the SOBPs, but the impact could become more pronounced if alternative (LET dependent) implementations are used. The resulting percentage change in the SOBPs was generally less than the percentage change in the parameters. The patient case showed similar effects as with the SOBPs in water, and parameter variations had similar impact on the biological dose when using the clinical MKM and the general MKM. The clinical LEM calculated the highest biological doses to both tumor and surrounding healthy tissues, with a median target dose ( D 50% ) of 40.5 Gy(RBE), while the MKM with HSG and V79 parameters resulted in a D 50% of 34.2 and 36.9 Gy(RBE), respectively. In all, the observed change in biological dose distribution due to parameter variations demonstrates the importance of accurate input parameters when applying the MKM in treatment planning.